Pinned post
Very proud to share our @iclr_conf paper: TopoNets! High-performing vision and language models with brain-like topography! Expertly led by grad student @mayukh091 and @MainakDeb19! A brief thread..
Apurva Ratan Murty
@apurvaratan
Neuroscientist. Assistant Professor of Cognition and Brain Science @GeorgiaTech
1.2K Following 1.9K Followers 787 Posts
Latest posts 5
Seems as good a time any to post our recent position paper. 😉 @paleunahu Now selected as an Oral at the Neurips workshop on Interp4discovery
Our position paper has been selected for Oral at @NeurIPS interp4discovery. We propose a framework for turning local interpretability findings into testable mechanistic explanations: openreview.net/forum?id=8Duo4…
Thanks to Chen Jiang @PsychBoyH @apurvaratan and Sashank Varma.View quoted post on X
Another update from our topography work from (high-schooler!) Navneet Prakash. Navneet trained path-integrating RNNs with TopoLoss and found that they recover a much more brain-like center-surround RF connectivity structure than baseline models (w/o any post-hoc sorting)!
Today I want to share an update from our topographic ANN work led by @haidertahan. What does adding topography actually do to a model’s internal representations? Does it make is more brain like? Introducing TopoAudio
Excited to first share @alishdipani 's study. Alish and I had a simple q: Are category-selective units in ANNs actually like category-selective regions in the brain? (why is this imp? read until the end). Turns out this q itself is highly under-constrained!
I’ve been pretty terrible at talking about our work online. I’m going to try to fix that! There are a bunch of things from the lab I’d love to share, so look out for some cool work you may not know about over the coming days! :)View quoted post on X